The influence of PRAME immunohistochemistry on the diagnosis, diagnostic confidence and treatment recommendation by dermatopathologists: A single institution prospective survey-based study

The precise differentiation of melanomas from melanocytic nevi is a persistent diagnostic challenge warranting continuing development of improved ancillary testing for equivocal cases since discordant diagnoses have been reported in as many as 25% of cases reviewed at tertiary referral centers. ¹ Preferentially expressed antigen in melanoma (PRAME), a biomarker expressed in melanomas but not widely in benign tissues, has shown promise in retrospective studies as an immunohistochemical (IHC) marker distinguishing melanoma from nevi. Diffuse nuclear staining was previously observed in 87% of metastatic and 83.2% of primary melanomas, with negative staining in 86.4% of nevi. ^2–3 PRAME gene expression is also utilized as a biomarker in uveal and cutaneous melanoma.^{2, 4–5} These promising results have increased the use of PRAME IHC in dermatopathology practice.^6–7

Therefore, we prospectively examined the influence of PRAME IHC on the histopathologic diagnosis, diagnostic confidence and recommended clinical management for challenging melanocytic neoplasms. After IRB approval, all melanocytic neoplasms from routine practice for which PRAME IHC was ordered from December 2019 through July 2020 at a single tertiary academic referral center were prospectively included. Three board-certified dermatopathologists prospectively completed surveys,^8–9 before and after review of PRAME IHC, documenting preferred diagnosis, diagnostic confidence and treatment recommendations. Surveys were excluded if incomplete or if IHC was unsuccessful. PRAME IHC staining pattern (positive, negative, variable, equivocal) was recorded as documented in the final pathology report. Diagnoses were binned into categories of benign, indeterminate, or malignant, and changes between these categories were noted. Diagnostic confidence was rated on a seven point scale, and the number of cases with increased (>0), decreased (<0) or unchanged confidence (=0) was calculated. Similarly, the changes in treatment recommendations were noted. Consultations were obtained in 24% (13/67) of cases.

The cases (N=70) included a variety of benign, indeterminate, and malignant diagnoses, with indeterminate pre-survey diagnoses, including atypical junctional melanocytic proliferation (AJMP), being the most common. After PRAME IHC, the number of indeterminate cases were reduced by 8.9%, with diagnostic changes (between categories of benign, indeterminate or malignant) made in 16.1% of cases. Diagnostic confidence increased in 74.6% (p <0.0001, Table 1) of cases and was observed across all staining patterns (Table 2). As expected, changes in diagnosis from indeterminate to malignant warranting more aggressive clinical management occurred most frequently in the case of positive PRAME staining (11/12 cases, Table 2). Downgrading of diagnoses with corresponding changes in management after negative or equivocal PRAME IHC was observed less frequently (Table 2).

Table 1:

Favored diagnosis, diagnostic confidence, and treatment recommendation before obtaining (pre-test) and after review (post-test) of PRAME IHC.¹ Pre-test and post-test columns assess percentages of cases within a category overall (not pairwise changes between pre- and post- test results for a given individual case).

Characteristic	Pre-test	Post-test	p value*
Diagnosis 2			0.6209
Benign	5/56 (8.9%)	6/56 (10.7%)
Indeterminate	35/56 (66.1%)	32/56 (57.1%)
Malignant (in situ or invasive, with or without an associated nevus)	15/56 (25.0%)	18/56 (32.1%)
Not stated	11 cases (excluded)
Confidence			<0.0001
Very unsure	0/67 (0%)	0/67 (0%)
Unsure	5/67 (7.5%)	0/67 (0%)
Somewhat unsure	19/67 (28.4%)	5/67 (7.5%)
Neutral	3/67 (4.5%)	3/67 (4.5%)
Somewhat confident	26/67 (38.8%)	22/67 (33.8%)
Confident	14/67 (20.9%)	31/67 (46.3%)
Very confident	0/67 (0%)	6/67 (9.0%)
Not stated	0 cases
Treatment recommendation 3			1.0000
No further treatment necessary OR	18/63 (28.6%)	18/63 (28.6%)
Close clinical surveillance
Excision with wide margin OR	45/63 (71.4%)	45/63 (71.4%)
Wide local following MIS or melanoma guidelines OR
Sentinel lymph node biopsy and/or evaluation for metastasis
Not stated	4 cases (excluded)

¹Excluded cases that did not have both pre- and post-test survey answers.

²Excluded cases that did not state preferred diagnosis.

³Excluded cases that did not state treatment recommendation.

^*Chi square test was used to obtain p values while Fisher’s Exact Test was used when one of the incidences was less than 5.

Table 2:

Changes in diagnosis, confidence and treatment recommendation stratified by PRAME immunohistochemistry staining pattern.¹

Characteristic		PRAME IHC Stain				p value*
Diagnosis change 2	Total	Positive	Variable Positive	Negative	Equivocal	0.0528
No change	46	20	8	14	4
Benign to malignant	0	0	0	0	0
Benign to indeterminate	0	0	0	0	0
Indeterminate to malignant	6	5	0	1	0
Malignant to benign	0	0	0	0	0
Malignant to indeterminate	2	0	0	0	2
Indeterminate to benign	1	0	0	1	0
Total	55	25	8	16	6
Confidence change 3						0.1352
No change	13	4	3	3	3
Increased	49	24	5	15	5
Decreased	3	0	0	3	0
Total	65	28	8	21	8
Treatment recommendation change 4						0.0034
No change	47	15	7	19	7
More aggressive	12	11	0	1	0
Less aggressive	2	0	0	1	1
Total	61	26	7	20	8

¹Excluded cases that did not have both pre- and post-test survey answers, as well as cases for which the PRAME staining pattern was not explicitly defined in the pathology report.

²Excluded cases that did not state preferred diagnosis.

³Excluded cases that did not state confidence.

⁴Excluded cases that did not state treatment recommendation.

^*Fisher’s Exact Test was used to obtain p values when incidence was less than 5.

This prospective survey-based study demonstrated that the number of indeterminate diagnoses decreased slightly, and that diagnostic confidence increased markedly after PRAME IHC. As expected and correlated with its biologic role, positive PRAME staining was associated with malignant diagnoses, and resulted in more aggressive treatment recommendations.¹⁰ Importantly, increasing confidence in equivocal cases without knowledge of increased diagnostic accuracy can be problematic, and communicating diagnostic ambiguity to clinicians is important in such cases. It is also important to note that rather than retrospectively studying the PRAME staining pattern in unequivocal entities, this study sought to examine the effect of PRAME IHC on the diagnostic decision making of the dermatopathologists prospectively in all equivocal cases with PRAME IHC. Our observations corroborate the finding of retrospective studies that PRAME positivity is correlated with malignant diagnoses not only for unequivocal neoplasms but also in the setting of challenging melanocytic tumors.⁷ Although final diagnosis does not rely upon a single ancillary test, this data suggests that PRAME IHC is a welcome addition to the armamentarium of dermatopathologists. Additional research will further define the utility and appropriate interpretation and use of PRAME IHC in a wide variety of challenging melanocytic neoplasms.

Data Availability Statement:

The data that support the findings of this study are available from the corresponding author upon reasonable request.